As the rapid development of electronic technologies and communication technologies has occurred, information exchange by using an electronic apparatus has become an indispensable part in the daily lives of the people. For example, people often use an electronic apparatus, such as a mobile phone, or a computer, etc. to receive and transmit text messages and emails, etc., which brings great convenience to the lives and works of the people. When people use an electronic apparatus to input information, the output of a candidate character or word becomes very important. If the electronic apparatus may know the content to be inputted by the user and correspondingly output the character or word the user desires, the input efficiency will be greatly improved and the user experience will be optimized.
Currently, when a user inputs a Chinese phrase with Pinyin input method, he first has to input the syllable, and then selects the Chinese candidate phrase he wants. In most cases, one syllable has many corresponding Chinese candidate phrases which are called homophones. As shown in FIG. 1, syllable “shi.shi” may correspond to phrases , or , etc. There will be much more candidate phrases when the user uses Jianpin input method (acronym or initial Pinyin). As shown in FIG. 2, “s.s” in Jianpin input method may correspond to phrases , or , etc. Since a Jianpin combination may have many matched phrases, the user sometimes has to page down to get the correct one in a long candidate input list. Thus, how to output a desired candidate phrase according to the input of a user has become the direction of research of the industry.
Currently, many word engines like Sogou and Baidu can automatically adjust word and phrase (the Chinese candidate character mentioned above) sequences by moving the most frequently used word forward, but they still do not handle the problem that a user tends to use different word databases in different conversation contents. They also provide various customized word cells (e.g. a group of character names in a fashion movie) for users to download, but the word cells are not so specialized so that user has to find out the one he really wants. It is a waste of time and is not intelligent.
In addition, many people have their own styles in inputting text messages or emails. For example, some people like to use network popular words, and some frequently use terminologies or idioms, etc. However, it is not easy for a user to get a right candidate phrase when he inputs a non-network popular word or a special word that came from terms, news, games, or movies, etc.